DO AMNESTY PROGRAMS
ENCOURAGE ILLEGAL IMMIGRATION?
EVIDENCE FROM IRCA
Pia M. Orrenius
Madeline Zavodny
Research Department
Working Paper 0103
FEDERAL RESERVE BANK
OF
DALLAS
Do Amnesty Programs Encourage Illegal Immigration? Evidence from IRCA*
Pia M. Orrenius
Research Department
Federal Reserve Bank of Dallas
2200 N. Pearl St.
Dallas, TX 75201
(214) 922-5747
[email protected]
Madeline Zavodny
Research Department
Federal Reserve Bank of Atlanta
1000 Peachtree St.
Atlanta, GA 30309
(404) 498-8977
[email protected]
October 2001
Abstract: This paper examines whether allowing certain undocumented immigrants to legalize
their status leads to additional illegal immigration. We focus on the effects of the 1986
Immigration Reform and Control Act, which granted amnesty to over 3 million undocumented
immigrants. We find that apprehensions of persons attempting to illegally cross the U.S.-Mexico
border declined immediately following passage of the law but returned to normal levels during
the period when illegal immigrants could file for amnesty and the years thereafter. Our findings
suggest that the amnesty program did not change long-run patterns of illegal immigration from
Mexico.
JEL classification: J61
Key words: undocumented immigrants, illegal aliens, amnesty
*
We thank Gordon Hanson for providing the data on Mexican economic conditions and border enforcement. We
also thank Alan Viard and Jason Saving. The views expressed here are those of the authors and do not necessarily
reflect those of the Federal Reserve Bank of Dallas, the Federal Reserve Bank of Atlanta or the Federal Reserve
System.
Do Amnesty Programs Encourage Illegal Immigration? Evidence from IRCA
At the time of its passage in 1986, the Immigration Reform and Control Act (IRCA)
marked the biggest change in U.S. immigration policy in decades. IRCA granted amnesty to
undocumented immigrants who met specific provisions, required employers to verify workers’
eligibility to work legally, and increased funding for the Border Patrol. Over 3 million
individuals, including 2.3 million Mexicans, were granted legal permanent resident status under
IRCA. Opponents of the law claimed that it would encourage future illegal immigration,
notwithstanding tougher border enforcement, because it set a precedent for granting amnesty
(e.g., Anderson, 1986).1 Such concerns arose again recently as, fifteen years later, another
amnesty is under consideration. President George W. Bush is considering granting legal
residency to some of the estimated 3 to 4 million undocumented Mexican immigrants currently
present in the United States, a proposal strongly endorsed by Mexican President Vicente Fox.
Previous research on the effect of IRCA on flows of undocumented immigrants has
reached mixed conclusions. Based on data gathered from seven Mexican communities in 198789, Donato, Durand and Massey (1992b) find little evidence that IRCA lowered the number of
undocumented Mexicans entering the United States. Based on estimates from decennial Census
and Current Population Survey data after correcting for changes in the size of the undocumented
immigrant population as a result of IRCA, Woodrow and Passell (1990) similarly conclude that
the annual change in the number of undocumented immigrants during 1986-88 was not
significantly different from changes prior to IRCA. Bean et al. (1990), in contrast, conclude
from Immigration and Naturalization Service (INS) data on apprehensions during 1977-89 that
1
For example, Congressman Lamar Smith (2000) said that, “Amnesty actually precipitates even more illegal
immigration, as individuals come to join their amnestied relatives or are encouraged in the belief that if they can
just elude the Border Patrol and stay underground for a few years, they will eventually get amnesty themselves.”
1
apprehensions declined by about 27 percent after IRCA; the effect did not change significantly
during the 35-month post-IRCA period they observe. White, Bean and Espenshade (1990) report
a similar result using INS apprehensions data for 1977-88.
This study reexamines the effect of IRCA on undocumented immigration flows to the
United States. Unlike previous studies, we distinguish between the period immediately
following passage of the bill and the period when amnesty applications were accepted as well as
assessing the overall effect of IRCA on illegal immigrant flows through 1996. Because the main
IRCA legalization program required that applicants be present in the United States, the law may
have encouraged illegal immigration immediately following its passage or during the application
period but not after the amnesty program expired. Expected tougher border enforcement in the
future also may have prompted undocumented immigration shortly after the law was enacted. In
addition, IRCA may have encouraged immigration even after the amnesty period expired if the
law fostered beliefs that other amnesty programs would occur in the future. Alternatively, the
long run effects of IRCA may differ from the short term effects as potential migrants learned
whether employment eligibility documentation requirements could easily be circumvented.
Establishing the effect of an amnesty program on unauthorized immigration is important
for several reasons. First, the number of illegal immigrants present in the United States is
substantial. An estimated 7 to 8.5 million immigrants—including 3.9 to 4.5 million Mexicans—
were illegally present in the United States in 2000, and the annual net flow of undocumented
immigrants is estimated at 275,000 (Porter, 2001; Immigration and Naturalization Service,
2000). It is generally believed that undocumented immigrants have lower skill levels, on
average, than other immigrants (Fix and Passel, 1994; Tienda and Singer, 1995); this is partly
because most undocumented immigrants are from Mexico and, more recently, other Latin
2
American countries, which have low average levels of education. The negative effects of
immigration on natives’ labor market outcomes are generally believed to be concentrated among
high school dropouts, so an additional influx of low-skilled immigrants could have adverse
effects on some natives (Borjas, Freeman and Katz, 1997).2 Larger flows of low-skilled
immigrants may also exacerbate income inequality and impose fiscal burdens on state and local
governments (Smith and Edmonston, 1997).
The next section details the provisions of IRCA. We then discuss the data and methods
used to examine whether the Act encouraged undocumented immigration. The results indicate
that the flow of undocumented immigrants slowed for six months after passage of IRCA but then
returned to previous levels. We do not find that IRCA stimulated illegal immigration in the
hopes of receiving amnesty, but we also do not find that the law discouraged undocumented
migration in the long run.
Legislative Background
IRCA included two separate legalization programs: the Legally Authorized Workers
(LAW) program and the Special Agricultural Workers (SAW) program. Under the LAW
program, illegal aliens who had lived in the United States since January 1, 1982, and met certain
other criteria, could become temporary legal residents. Successful applicants could then become
legal permanent residents (LPRs) after 18 months by meeting several criteria, such as
demonstrating basic knowledge of the English language and American civics. Almost 1.8
2
However, Bean, Lowell and Taylor (1988) suggest that undocumented immigrants are a complement to other
workers. Hanson, Robertson and Spilimbergo (1999) find little relationship between enforcement along the U.S.Mexico border and wages in U.S. border cities, suggesting that increased enforcement either does not affect
undocumented migrant flows or that wages are not responsive to changes in the number of undocumented
immigrants.
3
million applications were filed for the LAW program. The period for filing an amnesty
application under the LAW program was from May 5, 1987 until May 4, 1988. The SAW
program required that illegal immigrants have worked in U.S. agriculture for at least 90 days
during each of the previous three years or for at least 90 days during the last year to receive
temporary permanent resident status. SAWs could then receive LPR status in one or two years.
The SAW application period was from May 5, 1987 until November 30, 1988, six months longer
than the LAW program, and almost 1.3 million SAW applications were filed. SAW applicants
were not required to be residents of the United States to qualify for the program, only to have
met the agricultural work requirement. Over 2.7 million persons were granted amnesty under the
two programs, about 70 percent of whom were from Mexico.
A substantial number of applications for the IRCA amnesty program are believed to have
been fraudulent. For example, the entire qualifying foreign-born labor force, both legal and
illegal, believed to meet the SAW provisions was estimated at about 300,000, far less than the
more than 1 million SAW applications granted (Passel, 1999). Based on surveys conducted in
Mexico, Donato and Carter (1999) estimate that, in their samples, 73 percent of LAW
applications were fraudulent, and 40 percent of SAW applications. Cornelius (1989) reports that
28 percent of SAW applicants did not meet the program’s conditions in his sample of Mexicans.
Such widespread fraud suggests that people may have crossed the border after the bill’s passage
and then obtained fake documents indicating continuous residency in the United States since
January 1982 or agricultural work during the eligibility period.
Other IRCA provisions required employers to ask job applicants for documents
establishing legal status to work in the United States and imposed civil penalties (criminal on
repeat offenders) on employers who knowingly hire illegal aliens. The law also called for a 50
4
percent increase in personnel for the Border Patrol and increased the agency’s funding during
that fiscal year and the next by about 75 percent. This increase in Border Patrol personnel,
combined with legalization for millions of previously illegal aliens, would likely result in fewer
illegal border crossings and a lower number of undocumented immigrants in the United States.
The act also allowed persons who had resided illegally and continuously in the United States.
since January 1, 1972, to immediately qualify for LPR status.
The passage of IRCA was several years in the making. Immigration reform, including
amnesty for undocumented immigrants, was proposed in Congress in 1981, and the Senate
passed various bills in 1982, 1983 and 1985. The House of Representatives, however, was less
willing to back immigration reform, primarily because of concerns about the amnesty proposal.
On September 26, 1986, the House voted not to take up the immigration reform bill but then, in a
“stunning reversal,” passed the measure on October 9 (Congressional Quarterly, 1987: 65). The
bill was signed into law on November 6, 1986. Undocumented aliens are thus unlikely to have
anticipated passage of an amnesty program or to have illegally crossed the border before October
1986 in the hopes of an amnesty program being created. However, people may have responded
to the amnesty program after it was passed and entered the United States soon after passage of
the law intending to fraudulently file for amnesty.
There are several reasons the amnesty program may have encouraged illegal immigration.
Immigration quotas for non-immediate family members are binding for Mexicans, so many
family members of U.S. residents have to wait years for LPR (“green card”) status. An amnesty
program such as IRCA may provide a quicker route to becoming an LPR. For persons who are
not eligible for LPR status under the available categories (family, employment-based, refugee,
etc.), amnesty may offer the only means of obtaining a green card. Individuals may have entered
5
the United States without proper documents intending to file for LPR status under IRCA or in the
hopes of another amnesty program. IRCA also may have prompted some family members of
amnesty recipients to illegally migrate to the United States.
However, IRCA also contained several provisions that may have deterred undocumented
immigration. The law included a substantial increase in Border Patrol funds and personnel,
which would be expected to deter undocumented immigration by making potential immigrants
perceive that the costs of successfully crossing the border were higher or the likelihood of
success lower. The law also included sanctions for employers who knowingly hired
undocumented immigrants and required workers to produce documents verifying their
employment eligibility. The flow of undocumented immigrants would be expected to decline if
the law reduced the demand for illegal immigrant workers (and hence their wages) or if
immigrants believed it would be more difficult to obtain a job in the United States than was
previously the case.3
Data
We use INS data on the number of apprehensions at the U.S.-Mexico border as a proxy
for illegal immigration inflows.4 The number of apprehensions is, of course, not an ideal
measure of the number of illegal migrants successfully entering the United States or even of the
3
Several studies conclude that IRCA had negative effects on employment outcomes of undocumented immigrants
(Bansak and Raphael, 2001; Donato, Durand and Massey, 1992a; Donato and Massey, 1993). However, research
has not been able to distinguish whether such negative effects are due to a decline in demand or an increase in
supply of undocumented immigrants.
4
The INS data on apprehensions and enforcement are “linewatch,” or activities at the U.S. border instead of in the
interior. As discussed by Hanson and Spilimbergo (1999), the data include U.S. ports as well as borders and include
the Canadian border as well as the Mexican border. The vast majority (over 99 percent during 1977-96) of
linewatch apprehensions have occurred at the U.S.-Mexico border. According to Hanson and Spilimbergo,
linewatch activities at the U.S.-Mexico border are well correlated with total linewatch activities, but data on
linewatch activities at the Mexican border are only available beginning in 1977. We therefore follow Hanson and
Spilimbergo in using data on total linewatch activities.
6
number attempting to enter. In addition to counting the number of failed attempted crossings
instead of the number of successful crossings, the data include repeat apprehensions for the same
individual. The apprehensions data also do not reflect illegal aliens who enter legally and then
overstay their visas, who are believed to account for about one-half of illegal immigrants present
in the United States.
However, as noted by Bean et al. (1990), INS apprehensions data are believed to be
correlated with illegal crossings and are useful for examining periodic changes in the number of
such crossings. Espenshade (1995) concludes that the simple correlation between apprehensions
and the volume of illegal immigration is about 0.90 and that the flow of undocumented
immigrants is about 2.2 times the level of INS apprehensions. Espenshade cautions that the
exact magnitude of the effect of changes in variables on illegal immigrant flows cannot be
inferred from INS apprehensions data but that the direction of the effects can be identified; in
other words, we can determine the direction of the effect of IRCA on illegal immigration flows
from data on apprehensions but cannot identify the size of the change in the flow, only the size
of the change in apprehensions. The INS data used here are monthly for the period January 1969
through December 1996.
We use INS border patrol hours to measure border enforcement. The expected
relationship between enforcement and apprehensions is not certain because increased
enforcement could result in more apprehensions or could change migration behavior such that
apprehensions decline, either because fewer persons attempt to cross the border or because
migrants shift to areas less policed by the Border Patrol.
Figure 1 displays apprehensions and enforcement for the period 1969-96. Because there
are sizable seasonal swings in apprehensions, with apprehensions highest in the spring and
7
lowest in December, seasonally adjusted data for apprehensions are shown; the seasonal pattern
of apprehensions corresponds with the timing of demand for agricultural labor and with migrants
returning home for the winter holidays. Apprehensions generally increased from 1969-86 and
then declined through 1989. Apprehensions then increased through 1993 and spiked upward in
early 1995 and again in early 1996. Apprehensions appear to have declined during the period
between the passage of IRCA in November 1986 and the beginning of the filing period in May
1987. Apprehensions do not show a clear trend during the main 12-month filing period for
amnesty but appear higher than during the pre-filing period and higher than immediately after the
filing period ended. The descriptive statistics in Table 1 indicate that there was an average of
58,000 apprehensions per month during the pre-filing period, compared with almost 60,000 per
month during the 12-month filing period for amnesty.
Border Patrol man-hours spent enforcing the border also generally rose over time.
Enforcement hours rose during the IRCA pre-filing period but declined during the filing period
and immediately afterward until rising again in March 1988. Average monthly enforcement
hours declined from 237,000 during the pre-filing period to about 203,000 during the filing
period (Table 1).
Economic conditions in both the United States and Mexico are likely to affect the number
of persons attempting to cross the border.5 For the United States, the empirical model below
includes measures of the real average wage, the real federal minimum wage, and the average
unemployment rate.6 We deflate wages by the consumer price index (CPI) for urban consumers.
5
We do not include measures of economic conditions in other countries of origins because Mexicans account for the
vast majority of apprehensions, although the share of non-Mexicans apprehended at the border has increased over
time. Mexicans accounted for 96.1 percent of apprehensions during 1988-94 (Hanson and Spilimbergo, 1999).
6
The U.S. wage is a weighted average of hourly earnings for production workers in eight industries (mining,
construction, manufacturing, wholesale trade, retail trade, transportation and public utilities, finance/insurance/real
estate, and services). The weights are based on the industry distribution of Mexican-born men aged 15-64 in the
8
For Mexico, we include the real average manufacturing wage and the minimum wage, which are
deflated using the Mexican CPI; we do not include a measure of the Mexican unemployment rate
because reliable data are not available during the early part of the sample period.7 Because of the
importance of oil production to Mexico’s economy, we include the U.S. producer price index for
crude oil to capture changes in oil prices. We also include the real exchange rate between
Mexico and the United States, deflated using the CPI for both countries.
The number of visas issued is another factor that may influence the inflow of illegal
immigrants. The empirical model includes the number of persons from Mexico granted LPR
status (not including those granted LPR status as a result of IRCA) and the number of nonimmigrant visas issued to Mexicans. The number of new LPRs may influence illegal
immigration if people cross the border illegally to join relatives with legal status; alternatively,
people granted LPR status may have previously been entering the United States illegally, in
which case an increase in LPRs would be expected to lower apprehensions. Non-immigrant
visas may be a substitute for entering the United States without proper documents, so that lower
issuance of non-immigrant visas raises the number of apprehensions. The visa variables are
reported for fiscal years instead of monthly but are measured here as monthly averages within
fiscal years.
1960, 1970, 1980 and 1990 decennial Censuses who immigrated in the last five years and do not report being a
citizen (except for 1960, when citizenship was not asked).
7
In some years the Mexican manufacturing wage data include a December bonus of up to two or three times
monthly earnings (aguinaldo). We removed this bonus from the data by subtracting the average difference in
December wages between years with the bonus and years without the bonus from observations in years with the
bonus.
9
Methods
We regress border apprehensions on border enforcement, the measures of economic
conditions and immigration visas, and three IRCA policy variables. The IRCA variables are a
dummy variable for the pre-filing period from November 1986 to April 1987, a dummy variable
for the main IRCA filing period from May 1987 through April 1988, and a dummy variable for
the post-IRCA period beginning in May 1988. We use these three variables to measure the
effect of IRCA because the incentives for potential undocumented immigrants to enter the
United States may have differed between the period immediately after the program was
announced, the period when amnesty applications were accepted, and after the amnesty program
ended. Other results discussed below examine the robustness of the results to including variables
indicating when applications for only the SAW program were accepted and when immigration
reform bills with amnesty provisions—precursors to IRCA—had passed at least one body of the
U.S. Congress.
Because the apprehensions data display very high first-order serial correlation, we
estimate the regressions in first differences, a method that is robust to the presence of a unit root
in the disturbance term. All of the continuous variables are in log first differences (the fiscal
year variables are in log 12-month differences), and the regressions also include a linear time
trend and set of month dummy variables to control for the seasonal pattern of apprehensions.
We report both ordinary least squares (OLS) and instrumental variables (IV) results. The
IV regressions instrument for border enforcement, which is endogenous if increased flows of
illegal immigrants cause stepped-up border enforcement within the same month or if both border
crossings and enforcement respond to another factor not controlled for in the model. We
10
instrument for border enforcement with U.S. government expenditures on defense and with three
lags of border enforcement.
Results
Apprehensions clearly declined during the period immediately after IRCA was passed but
before the amnesty application filing period began. As Table 2 indicates, apprehensions were
about 11 percent lower during the pre-filing period than prior to IRCA’s passage, controlling for
other factors. Apprehensions during the filing period and thereafter were not significantly
different than before IRCA, according to the results. In results not shown here, we included
separate linear time trends for the pre- and post-IRCA periods. The estimated coefficients for the
trend variables were not significantly different, also indicating that IRCA did not substantially
affect illegal immigrant inflows.
Enforcement is positively associated with apprehensions in both the OLS and IV results,
with a 10 percent increase in enforcement associated with a 4.4 percent increase in
apprehensions in the OLS specification reported in column 1. As shown in Column 2, the
estimated magnitude increases with enforcement is instrumented with defense spending and
lagged enforcement; this result is surprising since endogeneity would be expected to bias the
estimated coefficient on the enforcement variable upward. Hanson and Spilimbergo (1999) also
find that the magnitude of the enforcement coefficient increases slightly when the variable is
instrumented.
Apprehensions decline when the real average manufacturing wage in Mexico rises, with a
10 percent increase in the wage lowering apprehensions by about 3.3 percent. This indicates that
improved economic conditions in Mexico reduce migration flows to the United States, a result
11
also reported by many other immigration studies. The real exchange rate is positively associated
with apprehensions at the 10 percent significance level, suggesting that more Mexicans illegally
migrate to the United States when the purchasing power of their dollar earnings rises in Mexico.
We do not find significant relationships between apprehensions and measures of economic
conditions in the United States, suggesting that “push” factors may play more of a role than
“pull” factors in undocumented Mexican immigration. INS admissions of LPRs or nonimmigrants do not appear to influence apprehensions.
The results in Table 2 are based on the period 1969-96. Previous studies examining the
effect of IRCA on apprehensions used a shorter time period, beginning in 1977 and ending in
either 1988 or 1989 (Bean et al., 1990; White, Bean and Espenshade, 1990). Table 3 shows the
results if our sample is restricted to January 1977 through December 1989. The post-IRCA
variable then is equal to one for the period May 1988 through December 1989 instead of, as in
Table 2, through December 1996. The results again indicate that apprehensions declined
immediately after IRCA’s passage but then returned to normal levels during the filing period and
thereafter. Previous studies that did not distinguish between the different phases of the IRCA
amnesty program but instead looked for differences at 12-month intervals found a negative effect
for the entire period after IRCA’s passage, whereas our results indicate that the negative effect
occurred only during the six months after the law was passed.
The above results do not indicate that apprehensions increased during the main IRCA
filing period from November 1987 to May 1988, when LAW and SAW applications could be
filed. The SAW amnesty program continued to accept applications for another six months after
the LAW program filing period ended in May, and over one-half of SAW applications were filed
during this period. It is generally believed that vast fraud occurred in the SAW program because
12
applications only had to submit documents indicating that they performed agricultural work in
the United States during the relevant period rather than documents proving continuous U.S.
residence since 1982, as required for the LAW program. The pattern of undocumented migration
during the months when only SAW applications could be filed therefore may have differed from
the pattern in the rest of the post-IRCA period.
To investigate this possibility, we included a dummy variable equal to one in May 1988
to November 1988 in the OLS and IV regressions (the post-IRCA dummy variable then equals
one starting in December 1988). The results, which are not shown here, do not indicate that
apprehensions were significantly different during the SAW-only filing period than during either
the pre-IRCA period or the post-IRCA period. As in the above results, apprehensions were
significantly lower during the pre-IRCA filing period than before IRCA was passed.
We also investigated whether illegal immigration flows appear to have responded to the
possibility of an amnesty program before IRCA was passed. An immigration reform measure
that included an amnesty program was passed by the Senate in 1982, 1983 and 1985 but did not
pass the House of Representatives in the same year. In all three years, the House Judiciary
Committee approved an immigration bill similar to that passed by the Senate earlier the same
year, but the legislation died when the House adjourned for the year without voting on it. In
1984, the House did pass an immigration reform bill that included an amnesty program, but
differences between the legislation and the bill passed by the Senate the previous year could not
be resolved. Despite the ultimate failure to pass a bill until late October 1986, potential migrants
may have believed that an amnesty program was in the works whenever at least one house of
Congress had passed an immigration reform bill.
13
We therefore created a variable equal to one when an immigration reform bill had passed
at least one body of the Congress and was not yet dead because the legislature had adjourned and
equal to zero otherwise. If people crossed the border in anticipation that an amnesty program
might be created in the future, this variable should be positively associated with apprehensions.
In results not shown here, however, the dummy variable indicating passage of an amnesty
program in at least one body of Congress is not significantly associated with increased
apprehensions. The pre-IRCA filing variable continues to be negatively associated with
apprehensions while the IRCA filing period and post-IRCA variables are not significantly
associated with apprehensions.
Conclusion
This study examined the effect of IRCA on undocumented immigration flows using data
on border apprehensions, a proxy for the number of people illegally entering the United States.
Apprehensions declined immediately after passage of IRCA but then returned to normal levels
during the amnesty filing period and thereafter. These results have several implications. Since
we find that apprehensions did not rise during the filing period, as would be expected if people
migrated to the United States to fraudulently apply for the program, it appears that amnesty
programs do not encourage illegal immigration. If anything, IRCA reduced the number of illegal
immigrants in the short run, perhaps because potential migrants thought that it would be more
difficult to cross the border or to get a job in the United States after the law was passed. An
amnesty program also does not appear to encourage illegal immigration in the long run in the
hopes of another amnesty program; we do not find a significant difference between
14
apprehensions after the IRCA amnesty program expired and before the program was created.
However, IRCA does not appear to have discouraged illegal immigration in the long run.
The INS apprehensions data used here do not allow us to distinguish between new and
returning illegal immigrants or to estimate the stock of illegal immigrants in the United States.
Our findings suggest that the net change in the stock of illegal immigrants was not affected by
IRCA in the long run. However, the net change in the stock of illegal immigrants could have
increased if illegal immigrants already present perceived that it was more difficult to cross the
border after the law was passed. Circular migration between the United States and Mexico could
have slowed, leading to a reduction in apprehensions of illegal immigrants who usually reside in
the United States but temporarily returned to Mexico to visit relatives and had to illegally recross
the border. A decline in apprehensions due to reduced circular migration could offset an increase
in apprehensions among new illegal immigrants, suggesting that the illegal immigrant population
could have risen. Future research should examine this possibility using individual-level data on
border crossing patterns among both repeat and first-time illegal migrants.
15
References
Anderson, Annelise, 1986. “Illegal Aliens and Employer Sanctions: Solving the Wrong
Problem.” Hoover Essays in Public Policy. Stanford, CA: Hoover Institution.
Bansak, Cynthia, and Steven Raphael, 2001. “Immigration Reform and the Earnings of Latino
Workers: Do Employers Sanctions Cause Discrimination?” Industrial and Labor Relations
Review 54 (January): 275-295.
Bean, Frank D., B. Lindsay Lowell, and Lowell J. Taylor, 1988. “Undocumented Mexican
Immigrants and the Earnings of Other Workers in the United States.” Demography 25
(February): 35-52.
Bean, Frank D., et al., 1990. “Post-IRCA Changes in the Volume and Composition of
Undocumented Migration to the United States.” Undocumented Migration to the United States,
ed. Frank D, Bean, Barry Edmonston, and Jeffrey S. Passel, Santa Monica, CA: RAND, pp. 111158.
Borjas, George J., Richard B. Freeman, and Lawrence F. Katz, 1997. “How Much Do
Immigration and Trade Affect Labor Market Outcomes?” Brookings Papers on Economic
Activity, 1997 (1): 1-90.
Cornelius, Wayne A., 1989. “Impacts of the 1986 US Immigration Law on Emigration from
Rural Mexican Sending Communities.” Population and Development Review 15 (December):
689-705.
Congressional Quarterly, 1987. 1986 CQ Almanac. Washington, DC: Congressional Quarterly.
Donato, Katharine M., and Rebecca S. Carter, 1999. “Mexico and U.S. Policy on Illegal
Immigration: A Fifty-Year Retrospective.” Illegal Immigration in America, ed. David W. Haines
and Karen E. Rosenblum, Westport, CT: Greenwood Press, pp. 112-129.
Donato, Katharine M., and Douglas S. Massey, 1993. “Effect of the Immigration Reform and
Control Act on the Wages of Mexican Migrants.” Social Science Quarterly 74 (September): 523541.
Donato, Katharine M., Jorge Durand, and Douglas S. Massey, 1992a. “Changing Conditions in
the US Labor Market.” Population Research and Policy Review 11: 93-115.
Donato, Katharine M., Jorge Durand, and Douglas S. Massey, 1992b. “Stemming the Tide?
Assessing the Deterrent Effects of the Immigration Reform and Control Act.” Demography 29
(May): 139-157.
Espenshade, Thomas J., 1995. “Using INS Border Apprehension Data to Measure the Flow of
Undocumented Migrants Crossing the U.S.-Mexico Frontier.” International Migration Review
29 (Summer): 545-565.
16
Fix, Michael, and Jeffrey S. Passel, 1994. Immigration and Immigrants: Setting the Record
Straight. Washington, DC: Urban Institute.
Hanson, Gordon H., Raymond Robertson, and Antonio Spilimbergo, 1999. “Does Border
Enforcement Protect U.S. Workers from Illegal Immigration?” NBER working paper no. 7054
(March).
Hanson, Gordon H., and Antonio Spilimbergo, 1999. “Illegal Immigration, Border Enforcement,
and Relative Wages: Evidence from Apprehensions at the U.S.-Mexico Border.” American
Economic Review 89 (December): 1337-1357.
Immigration and Naturalization Service, 2000. 1998 Statistical Yearbook of the Immigration and
Naturalization Service. Washington, DC: Government Printing Office.
Passel, Jeffrey S., 1999. “Undocumented Immigration to the United States: Numbers, Trends,
and Characteristics.” Illegal Immigration in America, ed. David W. Haines and Karen E.
Rosenblum, Westport, CT: Greenwood Press, pp. 27-111.
Porter, Eduardo, 2001. “Illegal Immigrants May Total 8.5 Million.” Wall Street Journal, August
14, p. A4.
Smith, James P., and Barry Edmonston, eds., 1997. The New Americans, Washington, DC:
National Academy Press.
Smith, Lamar, 2000. Lamar Smith Press Conference Statement on Amnesty, October 12.
Available at http://wwwa.house.gov/lamarsmith/im-101200.html
Tienda, Marta, and Audrey Singer, 1995. “Wage Mobility of Undocumented Workers in the
United States.” International Migration Review 29 (Spring): 112-138.
White, Michael J., Frank D. Bean, and Thomas J. Espenshade, 1990. “The U.S. 1986
Immigration Reform and Control Act and Undocumented Immigration to the United States.”
Population Research and Policy Review 9 (May): 93-116.
17
Table 1
Sample Means, by Time Period
Pre IRCA
1/69-10/86
IRCA Pre-Filing IRCA Filing
11/86–4/87
5/87-4/88
Apprehensions
(in thousands)
Enforcement
(hours, in thousands)
U.S. wage
33.31
(1.48)
147.45
(2.68)
9.71
(.03)
U.S. minimum wage
3.74
(.02)
U.S. unemployment rate
6.75
(.11)
Mexican wage
110.33
(.96)
Mexican minimum wage
19.75
(.23)
Oil price index
50.34
(2.23)
Real exchange rate
3.81
(pesos per $)
(.05)
Legal permanent residents
5.19
admitted from Mexico (thousands)
(.08)
Non-immigrant visas issued
119.67
to Mexicans (thousands)
(2.84)
Number of observations
214
57.93
(4.66)
237.37
(4.46)
9.16
(.02)
3.00
(.01)
6.60
(.08)
82.43
(1.43)
13.04
(.28)
48.70
(2.17)
6.06
(.03)
6.03
(.00)
74.14
(.00)
6
Note: Standard errors are in parentheses.
18
59.84
(4.57)
203.03
(4.78)
9.04
(.01)
2.91
(.01)
5.88
(.07)
82.36
(1.01)
12.14
(.23)
54.46
(1.29)
5.40
(.13)
7.13
(.28)
79.12
(1.27)
12
Post IRCA
5/89-12/96
65.07
(2.09)
259.53
(5.58)
8.61
(.02)
2.87
(.01)
6.10
(.08)
94.73
(1.14)
8.64
(.12)
56.45
(1.15)
3.99
(.06)
7.75
(.27)
114.79
(1.48)
104
Table 2
First-Difference Regression Results for Apprehensions at the U.S.-Mexico Border, 1969-96
IRCA pre-filing period
IRCA filing period
Post-IRCA period
Enforcement
U.S. wage
U.S. minimum wage
U.S. unemployment rate
Mexican wage
Mexican minimum wage
Oil price index
Real exchange rate
Legal permanent residents
admitted from Mexico
Non-immigrant visas issued
to Mexicans
Trend
Adjusted R2
Number of observations
(1)
OLS
(2)
IV
-.112*
(.048)
.022
(.036)
.020
(.027)
.441**
(.086)
-.689
(1.941)
-.499
(.309)
-.165
(.211)
-.332*
(.144)
-.055
(.131)
.046
(.094)
.205
(.123)
-.019
(.020)
-.004
(.029)
-.002
(.002)
.772
335
-.112*
(.048)
.024
(.037)
.020
(.027)
.565*
(.230)
-.710
(1.948)
-.492
(.310)
-.208
(.224)
-.350*
(.148)
-.046
(.133)
.047
(.095)
.209
(.124)
-.018
(.020)
-.003
(.030)
-.002
(.002)
.771
335
* p<.05; ** p<.01
Note: The time period is January 1969 to December 1996. All continuous variables except the time trend are in log
first differences (fiscal year variables are 12-month differences). Month dummy variables are also included. In
column (2), enforcement is instrumented with defense spending and three lags of enforcement. Standard errors are
in parentheses.
19
Table 3
First-Difference Regression Results for Apprehensions at the U.S.-Mexico Border, 1977-89
IRCA pre-filing period
IRCA filing period
Post-IRCA period
Enforcement
U.S. wage
U.S. minimum wage
U.S. unemployment rate
Mexican wage
Mexican minimum wage
Oil price index
Real exchange rate
Legal permanent residents
admitted from Mexico
Non-immigrant visas issued
to Mexicans
Trend
Adjusted R2
Number of observations
(1)
OLS
(2)
IV
-.125*
(.052)
.006
(.041)
-.006
(.039)
.387
(.210)
-1.983
(3.221)
-.194
(.675)
-.311
(.328)
-.371
(.217)
-.174
(.214)
.173
(.192)
.434*
(.186)
-.008
(.026)
-.015
(.035)
.002
(.004)
.768
156
-.134*
(.056)
-.014
(.047)
-.003
(.042)
-.475
(.726)
-1.390
(3.456)
-.069
(.728)
-.037
(.049)
-.160
(.285)
-.390
(.286)
.230
(.210)
.380
(.203)
-.016
(.028)
-.017
(.037)
.002
(.004)
.737
156
* p<.05; ** p<.01
Note: The time period is January 1977 to December 1989. All continuous variables except the time trend are in log
first differences (fiscal year variables are 12-month differences). Month dummy variables are also included. In
column (2), enforcement is instrumented with defense spending and three lags of enforcement. Standard errors are
in parentheses.
20
Figure 1. Border Apprehensions and Enforcement, 1969-96
140000
420000
IRCA filing period
360000
IRCA pre-filing period
100000
300000
80000
240000
60000
180000
40000
120000
20000
60000
0
1969
0
1971
1973
1975
1977
1979
1981
1983
1985
1987
1989
Year
Apprehensions
Enforcement
1991
1993
1995
Enforcement Hours
Apprehensions
120000